# Uncovering the molecular targets of phytocannabinoids: mechanistic insights from inverse molecular docking fingerprint approaches

## TL;DR

This study uses a computational method to uncover new protein targets for minor cannabinoids, suggesting potential therapeutic uses without the side effects of THC.

## Contribution

A novel inverse molecular docking fingerprint method with hierarchical clustering is introduced to analyze cannabinoid binding patterns.

## Key findings

- Minor cannabinoids show high docking scores with proteins linked to cancer, neurological disorders, and inflammation.
- THC-class and CBD-class cannabinoids form distinct clusters based on their binding patterns.
- CBL-class cannabinoids exhibit binding patterns similar to THC-class, suggesting therapeutic potential without psychotropic effects.

## Abstract

Among diverse chemical profile of Cannabis sativa L., over 100 phytocannabinoids have been identified. The major cannabinoids
Δ
-9-THC and CBD are well-studied, with approved palliative and therapeutic applications such as appetite stimulation, antiemetic therapy, pain management and epilepsy treatment. However,
Δ
-9-THC’s psychotropic effects limit its broader use. Minor cannabinoids exhibit therapeutic promise for a variety of conditions, potentially offering therapeutic potential without the adverse effects of
Δ
-9-THC.

We explored 14 cannabinoids with an inverse molecular docking approach, docking each cannabinoid into
>50000
human protein structures from the ProBiS-Dock database. We validated our inverse molecular docking protocol using retrospective metrics (ROC AUC, BEDROC, RIE, enrichment factors, total gain). We apply the novel inverse molecular docking fingerprint method to better analyze the binding patterns of different cannabinoids and extend the methodology to include hierarchical clustering of fingerprints.

Our analysis of the inverse molecular docking results identified high scoring targets with potential as novel protein targets for minor cannabinoids, the majority associated with cancer, while others have connections with neurological disorders and inflammation. We highlighted GTPase KRas and hematopoietic cell kinase (HCK) as very promising potential targets due to favorable docking scores with almost all investigated cannabinoids. We also find multiple matrix metalloproteinases among the top targets, suggesting possible novel therapeutic opportunities in rheumatic diseases. An analysis of inverse molecular docking fingerprints shows similar binding patterns for cannabinoids with similar structures, minor structural differences still suffice to change the affinity to specific targets. Hierarchical clustering of inverse molecular docking fingerprints revealed two main clusters in protein binding pattern similarity, the first encompassing THC-class and similar cannabinoids, as well as CBL-class cannabinoids, while the second contained CBD, CBC, and CBG-class cannabinoids. Notably, CBL-class cannabinoids exhibited binding patterns more similar to THC-class cannabinoids than their CBC-class precursors, possibly offering potential therapeutic benefits akin to THC with fewer psychotropic effects.

This study highlights the therapeutic potential of minor cannabinoids and identifies their potential novel protein targets. Moreover, we demonstrate the utility of inverse molecular docking fingerprinting with clustering to identify compounds with similar binding patterns as well as identify pharmacophore-related compounds in a structurally agnostic manner, paving the way for future drug discovery and development.